What solutions, if any, are there for visible light transmission (VLC)?

Question

I want to know what ways are there to transmit data using LED's that are available to us today.

I currently can find only one reasonable resource for this topic, which is this: http://www.siemens.com/innovation/en/news_events/ct_pressreleases/e_research_news/2010/e_22_resnews_1002_1.htm

Is there any other information about this or resources I can use? What is the highest frequency can you blink a under $5 led at? What about a photo diode or photo transistor that can respond at such frequencies for under $5?

My aims are to have anything above 10KB/s but under $5 or so, with range that includes line of sight and it would be no more than one or two feet away. Am I just going at this entirely wrong and should instead use a RF solution? I am gearing towards visible light because it seems cheaper, does not give me the headaches of RF, and most certainly looks far more interesting than a RF or IR design, since you can actually see the process. :P

Answer 1

A project from the Czech Repuplic called RONJA gets 10 Mbps through more than a kilometer of open air with ordinary red LEDs. It uses a specialized amplifier design to drive the LED, but if all you want is to modulate an LED at a few 10s of kHz, then just wiring it to an output pin of a microcontroller will work.

Answer 2

I don't know what you think seeing the light will do for you. The LED will be flashing on and off so rapidly that it will just look like it's lit half way.

I don't know about visible LEDs, but it is possible to module IR LEDs quite fast. The right kind of photodiodes can also be fast. I've done bare IR LED to photodiode transmission at about 1 Mbit/second. The hardest part is the photodiode amplifier circuit. Photodiode output, especially when you need a few MHz bandwidth, is quite low. This needs to be amplified by 100 or more, but with a bandwidth of around 10 MHz if you want 1 Mbit/s data rate.

At these high data rates the decoding generally requires hardware. Normally I'd like to use manchester encoding for something like this, but unless you've got a piece of a FPGA you can implement it in, that's difficult to decode.

I ended up driving the IR LED from a UART and feeding the output of the receiver data slicer into a UART. The problem with that is that the average level is not 1/2 like with manchester, and is quite data dependent. That makes the data slicer trickier. I defined the protocol so that the transmitter was guaranteed to send bytes within some maximum time limit. If the transmitter had nothing to send, it would send a single NULL byte, and this was taken care of in the higher protocol. The receiver had to keep both the recent high and low received level, then slice at the average of those. The NULL byte kept the high level refreshed enough for the data slicer to be ready.

Answer 3

You can certainly blink visible LEDs very quickly-- there's not much difference between visible LEDs and IR LEDs and phototransistors. Data rates well over 1 MB/s are straightforward with low cost LEDs and phototransistors. So, there aren't any barriers to what you want to do.

You can also buy a laser pointer under $5 these days. Laser diodes are (loosely speaking) LEDs with a different internal architecture. And those can be modulated at extremely high rates, such that transmitting data at GB/s data rates is straightforward. However, receiving data at high rates is not nearly so simple.

Answer 4

The main problem with visible light is the simple fact that there is so much of it around. An LED at 100 yards can become quite unnoticeable in the background of optical noise. That is why infra red is used almost exclusively for communication. IR LEDs are much much brighter than the background IR radiation from heat (unless you live in a volcano), so a detector can be so much simpler.

The main exception is lasers. Because they can be targeted the detector is pretty much ignoring the background light and only seeing the laser light.

The only way to really use visible light without an eye burning laser is to use some form of optical guide for the light to travel down - something like fiber optics.

Answer 5

Very old question, I know, but it seems that none of the answers here mention TOSLINK, a well-established (standardized in 1983, apparently) system for data transmission using visible-light LEDs and photosensors. Its original specification gave it a limit of 3.1 Mbit/s, but modern implementations manage over 100 Mbit/s.

It uses a red LED emitter into a plastic optical fiber (quite a thick optical fiber too; this was way, way before we had the nice thin single-mode optical fiber that was optimized for data transmission), and some kind of detector on the other end--probably a CdS or CdSe photocell in the earliest incarnations (perhaps explaining the 3.1 Mbps limit?), but a photodiode or phototransistor in anything modern. Because it's a consumer standard (that's what the optical digital audio input on your TV is!), the cables for it are quite cheap, and you can probably get a cable, a transmit LED, and a phototransistor for a receiver for less than $5 total if you're not picky about quality.

(As an aside, just because it's an interesting little factoid, there really was no technical reason for TOSLINK to exist. The same signal it was made for (called S/PDIF, but more often known as just "digital audio") could be and was transmitted over copper wires. As far as I can tell, the only reason it exists is because some engineers thought it would be really cool. And it was!)

---

Since I noticed after typing the above that the question is asking about line-of-sight communication, which I read as implying they want free-space communication without optical fiber, I'll also mention your standard remote control.

Remote controls as used for televisions and other AV equipment use infrared LEDs and sensors, but there's no reason they can't use visible light. Contrary to what one of the other answers on here says, it's not the intensity that the receiver uses to distinguish the signal from the noise; otherwise an incandescent light bulb would easily drown out the signal. Rather, it uses modulation to distinguish signal from noise; the remote control contains an LED that alternates not between on and off, but between flickering at 38 kHz (most commonly; it may be a different frequency for different devices) and off. The receiver ignores anything that isn't at that specific frequency, and the data to be transmitted is modulated on top of that by switching the flickering on and off--not by changing the frequency, which is fixed. Since it's unlikely that anything else will be emitting at that specific frequency, it can know that the 38 kHz bursts are a signal from a remote control.

This entire system could also be easily implemented with visible light. If you want 10 kbps transmission, you may need to use a higher carrier frequency than 38 kHz, since that's not too much more than the frequency of your signal. But even just a few hundred kHz would be more than enough for reliable data transmission--as long as you don't have PWM-dimmed LED lighting that does its PWM at the same frequency, anyway!